In recent years, an input method of touching a touch panel disposed on a screen is widely used as a human-machine interface (HMI) between a human and a machine such as a mobile terminal or an in-vehicle instrument. The input method based on a touch panel provides improved operational feeling as compared to a conventional input method of operating, for example, switches. However, the input method based on a touch panel is similar to the conventional input method of operating switches in that the hands of a user need to be moved for operating purposes. Hence, the input method based on a touch panel is not good enough to reduce physical burden on the user. Further, the input method based on a touch panel makes it difficult to perform an operation in a situation where both hands of the user are used to drive a vehicle. Under these circumstances, a visual line input technology is proposed, for instance, in Patent Literature 1. The visual line input technology inputs information by detecting the visual line direction of the user and specifying an object selected by the user on the basis of the detected visual line direction. In addition, a technology proposed, for instance, in Patent Literature 2 applies the visual line input technology to a pointing device that coordinates with an image displayed on a head-up display, which presents information ahead of a driver of a vehicle.
A well-known method of detecting the visual line direction includes capturing an image of the pupils of a target with a camera while the pupils of the target are irradiated with infrared light, and estimating the visual line direction on the basis of an image formed by the infrared light reflected from the pupils in the captured image. This method detects the visual line direction by analyzing the status of the target's pupils in image processing. Therefore, a high-resolution image needs to be captured.
However, when detecting the visual line direction of a driver, it is necessary to capture an image of a wide range in order to properly capture the pupils of the driver. This is because the positions of the pupils vary with the physique and posture of the driver. Thus, it is necessary to use a high-performance camera capable of capturing a high-resolution image of a wide range. Besides, a considerably high arithmetic processing capability is required for rapidly processing a high-resolution image of a wide range.